Nicer imports

Hi, I wonder if we could have

import
  Data
    A
    B (ba, bb, bc)
    C  as C
    C1 as C1 (ca,cb,cc)
    D qualified
    E as E qualified
    F as F qualified (fa, fb, fc)
  Libs.Other

--  Data.E qualified as E -- both valid
--  Data.F qualified as F (fa, fb, fc)

to desugar to

import           Data.A
import           Data.B (ba, bb, bc)
import           Data.C as C
import qualified Data.D
import qualified Data.E as E
import qualified Data.F as F (fa, fb, fc)
import Libs.Other

so we could make syntax more dry? What it does if there is indentation it uses previous levels as prefix

Trac metadata

Trac field	Value
Version	7.11
Type	FeatureRequest
TypeOfFailure	OtherFailure
Priority	lowest
Resolution	Unresolved
Component	Compiler (Parser)
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system
Architecture

changed weight to 0

added Tfeature request Trac import parser labels

In your example, how would you distinguish whether or not you intend to import the module Data? Or more realistically, how do I import both Data.ByteString and Data.ByteString.Lazy without repeating ByteString?

import
  Data.ByteString
      @
      Lazy
import
  Data.ByteString {@, Lazy}

import Data.ByteString {@, Lazy}

You got me, there needs to be a symbol for indicating current node, and not sure which would be the best one @ here is just an example. Curly brackets to de-sugaring to indentation would help as well :)

changed milestone to %⊥

Unfortunately, this proposal is completely incompatible with how layout rule works in Haskell today, which is that is triggered on a few special keywords. There is nothing of the sort here, so you would have to enforce layout through an alternate mechanism. I think it's unlikely that we'd accept this proposal.

closed

Try to get support for this on one of the mailing lists first. See also: WorkingConventions/AddingFeatures.

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedWon'tFix

added Plow label